Protein-Coding Gene Repertoires

Abstract

Three previously unavailable prerequisites for gene repertoire analyses are established and used within this thesis: (1) a new tool (COGNATE, written in Perl) was developed and used that records all gene structure parameters instead of relying on summary metrics; (2) it was ensured that automatically generated gene models are suitable for gene structure analyses; and (3) a unique species sample was employed, which representatively covers a younger radiation and allows the comparison of holometabolous and hemimetabolous.<br /> Goal of this thesis was to describe and analyze the variation of gene structures within and between repertoires of the species-rich and diverse insects as a step towards understanding and explaining the mechanisms and driving factors of insect genome evolution.<br /> COGNATE was employed to evaluate the magnitude of changes in predicted structural properties of protein-coding genes due to manual curation by comparing annotated gene sets from seven insect species sequenced by the i5k initiative. The properties of automatically generated gene models and their manually curated replacements do not differ extensively, and major correlative trends regarding gene structures can be recovered from both sets. From these results I conclude that gene models yielded from unsupervised annotation procedures are a suitable data basis to characterize structural gene features of a whole repertoire. <br /> The gene repertoires of twelve Hymenoptera were structurally characterized using COGNATE. The two focal species, non-apocritan "symphytans", possess small genomes and gene repertoires, but a strikingly high GC content of more than 41 %. Striking features are highlighted and lead to the conclusion that structural analyses of gene structure parameters can be used to direct efforts of detailed investigations, for example focused on here discovered miniature introns of ants.<br /> The established analysis approach was used to assess the variability in protein-coding gene repertoires in a characterization of the repertoires of a large, unique species sample. Previous research suggested universal patterns of conservation within a gene repertoire in relation to others by which it could be partitioned: highly conserved genes, genes moderately conserved and unevenly distributed across taxa, and lineage- or species-specific genes. My results show that considerable differences exist in gene structural parameters between and within these gene sets beyond previously described patterns. This work provides a solid baseline of expectations on insect gene structure variation as well as manifold investigative leads prompting further research.